Use case
Transcribing Insurance Recorded Statements for Adjusters and SIU
By Juris Compiler AI · Updated September 27, 2026
Upload the audio or video file of a recorded statement, and Juris Compiler AI returns a speaker-labeled, timestamped transcript you can search, correct, and question. A 15-minute statement costs $11.25 at $0.75 per minute, with no subscription.
The result is a working transcript for claim review. It is not sworn or certified, and you check it against your own copy of the recording.
What recorded statements are used for in claims work
Adjusters take recorded statements from insureds, claimants, and witnesses to capture each person's account of a loss while it is fresh. Special investigations unit (SIU) investigators may return to those statements when a claim is referred for a closer look.
A transcript lets you quote the statement in a file note, search it for a date or a vehicle, and give a colleague something to read instead of a call to replay.
A recorded statement is not a deposition
In a federal civil case, a deposition must be taken before an officer authorized to administer oaths, or a person the court appoints, under Federal Rule of Civil Procedure (FRCP) 28(a)(1). The officer opens with an on-the-record statement that includes administering the oath or affirmation (Rule 30(b)(5)(A)) and later certifies the deposition in writing (Rule 30(f)(1)). State courts have their own rules, and they vary.
A recorded statement taken by an adjuster or investigator generally involves none of that: no Rule 28 officer, and none of Rule 30's notice, oath, and certification steps. Whether and how a statement can be used later depends on the applicable rules of evidence, which is a question for counsel. An insurance examination under oath is a separate procedure from a recorded statement.
Juris Compiler cannot swear a witness or certify a transcript, so if a claim becomes a lawsuit and a witness is deposed, the certified record comes from the deposition officer, not from this app. Our guide explains what a court reporter provides that an AI transcript does not.
From recording to reviewable transcript
Upload a claim's statements as audio or video files, such as MP3, M4A, WAV, or MP4, up to 10 at once. They are processed one at a time, in order, with no guaranteed turnaround time.
- Speaker labels. Speakers are labeled A, B, C until you rename them, for example B to the claimant's name. The names carry into the PDF and drafts.
- Confidence flags. Every line shows a color-coded confidence score, and the least certain lines get a marker on the left edge. Start your review there, but a high score does not mean a line is right.
- Corrections with an audit trail. Fix any line in place. The original wording is kept, along with who changed it and when.
- Search. Search a statement's text, with matches highlighted.
Transcription is English only, and you cannot add your own terms to the recognizer, so check names, addresses, and policy and claim numbers against your file. The app has no audio or video player, and the uploaded file is deleted after transcription, so keep the original recording in the claim file to listen to any flagged line.
The demo shows the review screen with a fictional deposition. It does not show redaction, drafts, or export.
Checking a timeline within one statement
Many claim questions come down to sequence: when the insured left, where each car was, when the water was first noticed. Two tools help you test one statement's account against itself.
Ask the Record
Ask about the timeline in plain English, such as "What time does the claimant say they left the parking lot?" or "When did the insured first see water on the floor?" Answers do not reliably cite timestamps, so search the statement for the time or place an answer mentions and read that line yourself before it goes into a file note.
Contradiction Finder
Run the contradiction finder on the statement to flag passages in tension, such as a departure time that moves from 5:30 to 6:15 later in the same call. Each flag quotes both lines with speaker and timestamp, so you know where to listen in your own copy of the recording. A flag is a reason to look closer, not a finding that anyone misstated the loss.
Across the claim file
The finder works within one statement. When the insured's and the claimant's timelines differ, lining them up is your job: file every statement under one matter per claim, then search the matter for the same street, time, or vehicle to see what each person said about it.
Each transcript also gets an automatic case summary with legal issues, possible actions, and a low, medium, or high risk rating. It is framed under US federal law, and state-law context is added only if you regenerate it and choose one state or all states. None of that is a coverage, liability, or fraud determination. For how each of these features works, see step 7 of our guide to transcribing a recording.
Redaction, drafts, and exports
Sharing a redacted transcript
The Redact PII toggle masks identifiers the engine detects, such as names, phone numbers, dates of birth, driver's license numbers, and bank and card numbers. It works on a copy, so the original transcript is never changed. Know its limits:
- It covers the on-screen view and the PDF only. The SRT file, drafted documents, and Ask the Record answers are not redacted.
- Organization names are not masked.
- Detection can miss things. Read every page of a redacted PDF before you send it.
- There is no redacted version of the audio.
Drafted reports are generic templates
There is no insurance or SIU report template. The closest fit is the investigator report draft, a generic template with sections for Case Overview, Chronology, Key Findings, Persons of Interest, and Recommendations. The incident report draft (the Police report button) is a generic, police-style template. You cannot edit a draft in the app or aim it at a particular party or issue, and its Recommendations are AI suggestions, not a referral or claim decision. Copy a draft as text or print it, then rework it in your own format.
Export formats
The transcript exports as a PDF with numbered lines, timestamps, and speaker names, marked as AI-generated and not certified, or as an SRT caption file. There is no page:line pagination. A statement covered entirely by free minutes can be reviewed and its drafts printed, but its PDF, redacted PDF, and SRT exports need a paid transcript.
Vendor review and data handling
Juris Compiler does not claim any compliance certification, such as SOC 2 or HIPAA, so if your company's vendor review has questions, start with the security page, which lists each subprocessor, what it receives, and retention in days, or talk to a human before you upload a claim file. Deleting a transcript does not ask the transcription provider to delete its copy, which follows that provider's own terms.
What a claim file costs to transcribe
The price is $0.75 per minute, billed per file and rounded up to the whole minute, with a 1-minute minimum. Here is one claim file with three recorded statements uploaded together:
| Statement | Recorded length | Billed | Price |
|---|---|---|---|
| Insured | 11:40 | 12 min | $9 |
| Claimant | 17:10 | 18 min | $13.50 |
| Witness | 24:05 | 25 min | $18.75 |
| Claim file | 55 min | $41.25 |
Rounding each statement on its own adds less than a minute per file. Up to 10 files can go in one upload and one Stripe checkout, which shows the exact total before you confirm. The pricing page has a calculator for other lengths, and you can sign up with Google or Facebook for 30 free minutes, no credit card.
Frequently asked questions
Is a transcribed recorded statement the same as a deposition transcript?
No. In a federal civil case, a deposition is taken before an officer authorized to administer oaths or a person the court appoints, who swears the witness and certifies the record, and state rules vary. A recorded statement generally has none of those steps. Juris Compiler AI produces a working transcript for review, not a sworn or certified record. How a statement can be used later is a question for counsel.
Can several adjusters work the same claim file?
Not in a shared workspace. Each transcript and matter belongs to one account, and there are no team plans or invites. To pass a statement to a colleague, a supervisor, or counsel, export the PDF from a paid transcript, with Redact PII on if the reader should not see the identifiers it masks. A transcript processed entirely on free minutes cannot be exported.
What does a redacted transcript hide, and what can it miss?
With Redact PII on, identifiers the engine detects, such as names, phone numbers, and Social Security numbers, are masked on screen and in the exported PDF, including the case summary, speaker names, and caption. The SRT file, drafted documents, and Ask the Record answers are not redacted, and organization names are not masked. Detection can miss things, so read every page before you share it.
How is a claim file with several recorded statements billed?
Each statement is billed as its own file at $0.75 per minute, rounded up to the whole minute. For example, a claim file with statements of 11:40, 17:10, and 24:05 is billed as 12 + 18 + 25 = 55 minutes, or $41.25, because each file is rounded on its own. You can upload up to 10 statements at once and pay for them in one Stripe checkout.
Try it on one of your own recordings
Your first 30 minutes are free when you sign up with Google or Facebook, no credit card needed. After that it is $0.75 per minute, and Stripe's checkout page shows the exact amount before you pay.
Related guides
Juris Compiler AI is a workflow and productivity tool, not a law firm, and does not provide legal advice. Using the platform creates no attorney–client relationship. Automated transcripts and AI-generated analysis may contain errors and are not a substitute for professional judgment. You are solely responsible for reviewing and verifying all output and for any decisions or actions taken based on it.